Hongdong Li
Papers
6
Total Citations
80
H-Index
4
About
Hongdong Li is a leading researcher in computer vision and robotics, whose work bridges perception, prediction, and autonomous navigation. His core research areas include action anticipation, visual localization, and robust visual odometry. Li’s major contributions span from foundational calibration techniques to cutting-edge generative models. Notably, his 2019 work on “Action Anticipation by Predicting Future Dynamic Images” (55 citations) introduced a novel framework for forecasting human actions by generating future visual representations, a key enabler for proactive robotic systems. He has also advanced cross-view localization for outdoor robotics, proposing a view-consistent purification method (2023) that robustly matches onboard camera views with satellite imagery for precise self-localization. In visual navigation, Li developed a dense optical-flow algorithm with uncertainty estimation (2021) to enhance monocular SLAM for ground and aerial robots, directly addressing challenges in ego-motion estimation and obstacle avoidance. His earlier work on single-shot extrinsic calibration of RGB-D cameras (2013) remains a practical reference for mixed reality and robotics setups. With recent explorations into diffusion models for hand motion prediction (2024), Li continues to push the boundaries of predictive vision, making his research highly relevant for students and engineers building next-generation autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1Action Anticipation by Predicting Future Dynamic Images55 citations · 2019
- 2View Consistent Purification for Accurate Cross-View Localization7 citations · 2023
- 3
- 4
- 5Prompting Future Driven Diffusion Model for Hand Motion Prediction4 citations · 2024
- 6Video Local Pattern based Image Matching for Visual Mapping2 citations · 2006